Etoposide is used in combined therapy with cisplatin or carboplatin as first-line treatment for small cell lung cancer (SCLC). Etoposide pharmacokinetics is intra- and inter-individually variable, leading to unpredictable safety and efficacy. The intravenous formulation has been used more extensively as the variability is distinctive especially after oral administration. However, in comparison with intravenous administration, oral administration has numerous advantages for the patient as well as economics benefits. This was our reason for optimization of oral etoposide dose which is currently adjusted only to the body surface area. Our research was started with the review of the literature about factors which have an impact on bioavailability of etoposide and options for its improvement. The pharmacokinetics of etoposide is influenced by genetic (CYP450 and UGT enzymes, membrane ABC transporters), physiological (age, hepatocellular carcinoma, liver cirrhosis) and environmental factors (pharmacokinetic interactions, poor water solubility, chemical instability in physiological fluids). Options for improvement of etoposide bioavailability include use of concomitant drugs with pharmacokinetic interactions influencing etoposide bioavailability, development of drug delivery systems ensuring rapid drug dissolution in the upper gastrointestinal tract or delaying drug release to target the colon, pharmaceutical forms that improve dissolution rate (e.g. phospholipid complex self-emulsifying drug delivery systems, nanostructured lipid carriers), use of more water-soluble form of etoposide (etoposide phosphate) or estimation of individual pharmacokinetic parameters.
We decided to optimize etoposide dose using Bayesian estimation of individual pharmacokinetic parameters with population pharmacokinetic model. Data on etoposide concentrations after administration of known etoposide dose was needed for the development of population pharmacokinetic model. The performance of pharmacokinetic study and determination of etoposide plasma concentrations after known dose of etoposide were needed. The study was started with the development of protocol in which we defined intravenous etoposide administration in the first cycle and oral administration in the following cycles. The analytical method for determination of etoposide concentrations in plasma had been developed before start of the study. More than 600 samples from SCLC patients were routinely analysed using this method. In addition, the analytical method for determination of etoposide concentration in dried blood spot was developed. This method is more patient-friendly and enables simpler sample handling. Comparing etoposide plasma concentrations and etoposide concentrations in dried blood spot in some patients it was ascertained that etoposide concentrations in dried blood spot were comparable to etoposide plasma concentrations when haematocrit (Hct) [CPL=CDBS/(1-Hct)] was considered. The method for determination of etoposide concentration in dried blood spot can be an alternative for determination of etoposide plasma concentrations (Hypothesis No. 1). We wanted to check which factors (found in the literature) influence plasma etoposide concentrations. Selected polymorphisms in gene MDR1 which could theoretically influence etoposide pharmacokinetics were determined for this purpose. Additionally, patient’s demographic data, biochemical results and data on pharmacotherapy were obtained from hospital records. The influence of all these data was tested with population pharmacokinetic model developed in the NONMEM program. Among all tested variables etoposide pharmacokinetics was influenced by patient’s characteristics and concomitantly administered medicines, while the influence of genetic factors was not significant (Hypothesis No. 2). These factors were included as covariates into the final pharmacokinetic model. The exposure (AUC) and inter-subject and inter-cycle variability in etoposide’s AUC after both routes of administration were determined using the final model. In comparison with intravenous administration, oral etoposide administration showed worse options for effective treatment (lower percent of patients with AUC >254,8 mgh/L). The overall variability as well as inter-individual variability and between-cycle variability were higher after oral administration (Hypothesis No. 3). The existence of differences between oral and intravenous etoposide administration was examined. The adverse drug reactions in oral cycles were compared with intravenous cycles. Results of efficacy, i.e. response rate, progression free survival and overall survival in our study were compared with literature results after intravenous administration. Safety and efficacy did not differ between intravenous and oral administration (Hypothesis No. 4). However, results of our study showed that febrile neutropenia (FN) rates were in accordance with literature data (10-20%); however, the use of granulocyte colony-stimulating factors was relatively more frequent than recommended in guidelines. Unexpectedly high portion of neutropenia was ascertained after the first chemotherapy cycle, requiring frequent use of secondary G-CSF prophylaxis. The grade and/or frequency of neutropenia would have possibly been even higher if blood count had been additionally monitored at the time of neutrophil nadir. Furthermore, the relatedness of peak plasma concentration height with severity of neutropenia was observed (Hypothesis No. 5); concentrations were the highest in patients with FN and declined to the lowest levels observed in patients without neutropenia. It was shown that appropriately performed clinical study, development and use of appropriate method for determination of etoposide in plasma and development of appropriate pharmacokinetic model provide optimization of oral etoposide treatment (Hypothesis No. 6). From the aspect of pharmacokinetic variability and higher probability for effective treatment intravenous administration was better. However, in both intravenous and oral etoposide administration dose optimization according to patient's renal function and average bioavailability 40% is recommended. Oral treatment adjusted to bioavailability 50% is at least appropriate due to low etoposide exposure, inferior probability for effective treatment and high pharmacokinetic variability. Contribution to science:
Based on the results, it can be concluded that population pharmacokinetic model, which enables optimization of oral etoposide dose in individual patient, was successfully developed. This was achieved using a suitably and innovatively designed clinical study and with the development of a suitable analitical method with a sufficiently low limit of quantification, which enabled determination of etoposide plasma concentrations. In addition, analitical method DBS was developed and etoposide concentrations in dried blood spot were determined. In the case of etoposide, this method has not been developed yet. It has been proven that this patient and analist friendlier method can replace the method of etoposide plasma determination. Selected polimorphisms of gene which encodes MDR1 were determined in patients and their impact on etoposide pharmacokinetics was researched using population pharmacokinetic model. Using that, we contributed to an ever developing branch of science, pharmacogenetics. This allows us to reveal genetic properties of an individual and to enable the adjustment of pharmacotherapy for achiving as great safety and efficacy of treatment in individual patient as possible. Additional confirmation, that looking for possibilities for oral etoposide dose optimization in individual patient is a step in the right direction, is also an estimated surprisingly high rate of neutropenias, which indicates a need for greater activities for prevention of neutropenic events. The alternative to population pharmacokinetic model can be presented by the measurement of first day treatment etoposide plasma concentration (especially in case of existing patient-friendly method) and the decision on primary G-CSF prophylaxis. All of our research findings are/will be published in four scientific articles. It would be reasonable to continue our research with the validation of developed pharmacokinetic model in clinical practice and later use the same model in routine clinical practice. It is expected that approaches, tools, and methods of work presented will be applied also in the individualization of treatment using other drugs with similar pharmacokinetic properties.
|
